Toward Record of Scores and Stats: Count of the Stroke Number Using a Mobile Phone Accelerometer

This paper presents our method of recording golfers' score using an accelerometer embedded in a mobile phone. Our contribution is a new framework that can distinguish a target type of actions such as strokes from the other actions based on a classification of motion pattern into move and stay, and deep analysis is performed in the stay period. The stroke detection and reference vector collection processes are concurrently running to provide adaptability to the variation of phone's postures. Field trials show that the precision is 0.82 and recall is 0.89.